JP2001024987A - Successive scanning converting circuit - Google Patents

Abstract

PROBLEM TO BE SOLVED: To hold the image quality of motion compensation type successive scanline conversion by reducing the degradation of a displayed image even when a detected motion vector is incorrect. SOLUTION: An input video signal is respectively inputted to a motion vector detecting circuit 102, a motion compensation interpolating circuit 103 and an intra-field interpolating circuit 104. In the motion compensation interpolating circuit 103, the video signal of a preceding field, for example, is outputted after compensating the motion corresponding to the detected motion vector. Further, a vertical high frequency component is extracted from the motion compensated video signal of the preceding field by a vertical band pass filter 105. This extracted vertical high frequency component is composed with the vertical low frequency component of a current field provided from the intra-field interpolating circuit 104 by a composition circuit 106 and this composite signal is outputted as the interpolating signal of a scan line. Thus, even when the motion vector is incorrect, the quality of an image subjected to successive scanning conversion by a double speed converting circuit 107, can be maintained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a progressive scan conversion circuit for converting a 2: 1 interlace signal into a progressive scan signal, and more particularly to a progressive scan conversion circuit for generating an interpolation signal by detecting a motion vector.

[0002]

2. Description of the Related Art A progressive scan conversion circuit for converting a video signal of 2: 1 interlace scan into a progressive scan signal (non-interlace signal) has been developed in response to the demand for higher image quality of the current broadcast system (NTSC system). Have been.

In order to convert a 2: 1 interlaced signal into a sequential scanning signal, it is necessary to interpolate every other scanning line. Motion interpolation type processing and motion compensation type processing have been proposed for this interpolation processing. ing.

A motion-adaptive progressive scan conversion circuit inputs an input video signal and a video signal one frame before to a motion detection circuit, performs motion detection based on the difference between the two signals, and responds to the motion detection signal according to the motion detection signal. In the part, scanning line interpolation (referred to as inter-field interpolation) is performed by the signal one field before by the field delay circuit, and in the moving image part, the input video signal and 1
Scan line interpolation (called intra-field interpolation) is performed by obtaining an average value between adjacent scan lines from a signal one line before by a line delay circuit.

However, in such a motion-adaptive scan conversion circuit, in a moving image portion, since the scanning line interpolation is performed by the average value of the upper and lower adjacent scanning lines, the vertical resolution is not improved. In the stationary portion, the vertical resolution is improved because the scanning line is interpolated by the signal of the previous field. However, once the stationary portion starts moving, the portion is scanned by the average value of the adjacent scanning lines. Since interpolation is performed, the vertical resolution is deteriorated as compared with the stationary state, and there is a problem that a large difference causes unnaturalness.

On the other hand, a motion-compensated progressive scan conversion circuit interpolates a scanning line using a motion vector between frames and fields, and is a problem in the above-described motion adaptive progressive scan conversion circuit. Is effective in reducing the vertical resolution of moving images and interlacing of moving images (Japanese Patent Application Laid-Open No. 6-121289 describes the effect of interlace wrapping on image quality (interlacing interference)). Therefore, highly accurate motion vector detection is required.

[0007] The motion vector indicates from where on the previous screen a certain portion of the current screen has moved. There are several methods for detecting the motion vector, and the block matching method is often used because of the ease of processing by a digital circuit. In the block matching method, an image is divided into (M × N) pixels (for example, 8 pixels × 8 pixels).
The motion vector is detected on a block-by-block basis.

In the block matching method, as shown in FIG. 7, a target block composed of (M × N) pixels in a current frame or a current field is temporally different from the current frame or the current field. The degree of correlation with each candidate block including (M × N) pixels in the search range set on the reference frame or the reference field is evaluated. The evaluation value of the degree of correlation is obtained, for example, by cumulatively adding the absolute value of the difference between the pixel values at the same position in the target block and the candidate block. A displacement between a candidate block having the smallest evaluation value of the obtained correlation degree and a target block projected on a reference frame or a reference field is detected as a motion vector.

As described above, according to the block matching method, an optimal block most similar to the target block of the current frame or the current field (ie, having the highest degree of correlation) is extracted from within the search range of the reference frame or the reference field, and the motion vector is extracted. Is detected. Therefore, when a certain object is moving in parallel at a speed of V (the number of pixels / one field period) in one field period, the image of one field before is moved to the estimated moving amount V
If only the shift is performed, the block having the highest degree of correlation between fields becomes the block (that is, the pixel difference accumulated value between the target block and the candidate block between fields becomes the smallest).

However, in the case where the moving object is rotating or the object itself is deformed, it is difficult to accurately detect the motion vector, and the position where the pixel difference accumulated value in the block unit becomes the smallest is accurate. It cannot be said that it indicates a motion vector. The same applies to the case where the amount of motion or the direction differs between the background and the foreground.

If the detected motion vector is incorrect,
Since the image information having no correlation with the surrounding picture is interpolated in units of blocks, the image is deteriorated remarkably as a block-like failure.

FIG. 8 shows a conventional progressive scan conversion circuit. As shown in FIG. 8, a video signal of interlaced scanning is input to an input terminal 101, and a motion vector detection circuit 10
2. Input to the motion compensation interpolation circuit 103 and the intra-field interpolation circuit 104. The motion vector detection circuit 102 detects the motion of an image between fields as a motion vector MV. The motion compensation interpolation circuit 103 generates a video signal obtained by moving the previous field horizontally and vertically by the motion vector MV with respect to the input video signal of the interlaced scanning, and inputs the video signal to one input terminal of the synthesis circuit 106. The intra-field interpolation circuit 104
A signal of a scanning line to be interpolated is generated by an in-field signal processing for an input video signal of interlaced scanning.

In the synthesis circuit 106, the output signal of the motion compensation interpolation circuit 103 and the output signal of the intra-field interpolation circuit 104 are
The digital signal is added according to the motion vector MV, and further processed so as not to overflow or underflow the processing range of the digital signal. The input signal supplied to the input terminal 101 and the interpolation signal output from the synthesizing circuit 106 are input to the double-speed conversion circuit 107, and the double-speed converted sequential scanning signal is output to the output terminal 108.

Problems of the operation of the conventional example shown in FIG. 8 will be described with reference to FIG. FIG. 9 shows an example in which a circular object moves from left to right between fields with a square background.
A case where motion compensation interpolation is performed from the image of FIG. 9A to the image of FIG. 9B will be described. What is indicated by a broken line is a target block for searching for a motion vector. A circular object has moved from FIG. 9 (a) to FIG. 9 (b) in one field.
When the block shown by the broken line in FIG. 2B is searched in FIG. 2A, the arrangement shown by the broken line has the highest correlation. But,
What is problematic in this example is the shaded portion in FIG. 9 (b).

Since the hatched portion in FIG. 9B is interpolated based on the motion vector from FIG. 9A, the interpolated scanning line in the hatched portion in FIG. Is displayed (that is, a rectangular internal image), which degrades the image quality.

If the shaded area shown in FIG. 9B is large, it is determined that there is no correlation between the blocks at the detection stage of the motion vector detection circuit 102, and the synthesis circuit 106 performs the intra-field interpolation from the intra-field interpolation circuit 104. Selectively output signals. In this case, the vertical resolution cannot be improved for the circular portion because the signal is not the motion compensation interpolation signal from the motion compensation interpolation circuit 103. If the shaded area of (b) is small, the synthesis circuit 106 mainly selects and outputs the motion compensation interpolation signal by the motion compensation interpolation circuit 103, as described above.
Since the interpolation is performed from the broken line in (a), the image quality deteriorates in an area corresponding to the hatched portion in (b).

As a method for preventing image quality deterioration, Japanese Patent Application Laid-Open No. 4-257183 has been proposed. In this conventional example, as shown in FIG.
The median value of the pixel value b on the existing scan line L1, the pixel value b 'on the scan line L2 below the interpolated scan line, and the pixel value on the scan line by motion compensation is obtained. Or,
From the pair of pixel values (a, a '), (b, b'), and (c, c '), the pair that minimizes the difference between the two pixel values is (p, p'), and this pair (p, p ') ') Is used for median filtering instead of the pair (b, b').

Alternatively, it is determined whether the pair (p, p ') in which the difference between the two pixel values is minimum corresponds to the vertical direction, that is, whether (p, p') = (b, b '). If this condition is satisfied, the median value between the pixel values b and b 'and the interpolation value by motion compensation is obtained. If the condition is not satisfied, the average value of the pixel values p and p' is used as the interpolation value. Output.

However, in the conventional example shown in FIG. 10, in a signal in which a vertical high frequency component is folded back to a vertical low frequency by interlace scanning, the vertical high frequency component cannot be interpolated.
This will be described with reference to FIG.

FIG. 11A shows three fields f1 to f3.
When the black line is moving at a speed of 2 lines (2 lines / field) for 21 field periods between the lines, the positional relationship between pixels used to interpolate the scanning line at the X point of the field f2. Is shown as an example.
The positional relationship between the pixels shown here is shown in FIG.
This applies to the positional relationship of (c).

FIG. 11 (b) shows interpolation by conventional motion compensation, in which point A of field f1 and field f
3, a motion vector can be detected from the point B. At the point X, the pixel value YC is obtained as YC = YA or YC = (YA +
YB) / 2 is interpolated (YC is indicated by a black circle C). Note that YA and YB indicate pixel values at points A and B, respectively.

However, the conventional example described in FIG.
No. 257183), as shown in FIG.
2, the median value between the pixel values YD and YE at the points D and E and the motion compensation interpolation value YC based on the adjacent fields f1 and f3 is calculated and interpolated. The pixel value YD or YE is interpolated (because the pixel values are considered to be almost the same in the adjacent lines D and E in the same field, and the median value YC 'is indicated by a white circle C'). Therefore, in the example of FIG.
Since the interpolation is performed within the same field, no vertical high frequency component is compensated for, and the vertical resolution is not improved.

Here, an example in which the object is rotating is shown in FIG.
This will be described with reference to (a) to (c). That is, an example in which the object in the field (N) in FIG. 12A is rotated in the field (N + 1) as shown in FIG. In this case, when the motion vector detection by the block matching method is adopted, the field (N + 1) shown in FIG. 12B is replaced by 8 pixels × N on the field (N) shown in FIG.
The image at the position where the pixel difference accumulated value is the minimum in the block of 8 lines is moved to the field (N + 1) in block units and the scanning line interpolation is performed.
Field (N) and field (N +
The two images of 1) are displayed in an overlapping state, and significantly deteriorate the original image.

In the example shown in FIG. 12, since the image of FIG. 12B is rotated with respect to the image of FIG.
Are not matched even if the image is translated in blocks. In particular, in a case of a small difference in luminance between the foreground and the background, it is difficult to obtain a difference between blocks of a field of a moving object, so that there is a problem that interpolation between fields does not operate properly.

[0025]

As described above, a motion compensation type progressive scan conversion circuit requires a highly accurate motion vector. However, despite the large circuit size, detection errors are completely prevented. It is very difficult. In particular, with the block matching method, it is difficult to detect an accurate motion vector when the image is rotated or deformed, and the sequentially converted image may be significantly deteriorated. Further, in a pattern of a repetitive pattern, since a plurality of points having a high correlation are detected, there is a problem that the motion vector detection is easily erroneous.

Therefore, the present invention provides a method for determining whether the motion vector detected as described above is incorrect (for example, when the image is rotated or deformed, when the luminance difference between the foreground and background is small, It is an object of the present invention to provide a progressive scan conversion circuit that can reduce the deterioration of a displayed image even in the case of a pattern of a repetitive pattern) and maintain the image quality of the motion compensated progressive scan line conversion. .

[0027]

According to a first aspect of the present invention, there is provided a progressive scan conversion circuit comprising: a motion vector detecting circuit for detecting a motion of an image between frames or fields of a video signal as a motion vector; Accordingly, a motion compensation interpolator for motion-compensating and outputting a video signal of at least one adjacent field, a circuit for extracting a vertical high-frequency component of the motion-compensated field, and generating a vertical low-frequency component of the current field. And a combining circuit for adding the vertical high-frequency component and the vertical low-frequency component and outputting the result as an interpolation signal.

Further, the progressive scan conversion circuit according to the invention of claim 4 is adjacent to a motion vector detection circuit for detecting a motion of an image between frames or fields of a video signal as a motion vector in accordance with the detected motion vector. A motion compensation interpolator for motion-compensating and outputting the video signal of at least one field, a circuit for extracting a vertical high-frequency component of the motion-compensated field, and a vertical low-frequency component based on a correlation determination between lines in the current field. A circuit for generating a band component; and a synthesizing circuit for adding the vertical high band component and the previous low vertical band component and outputting the result as an interpolation signal.

According to the progressive scan conversion circuit of the first and fourth aspects of the present invention, the synthesized signal of the vertical high frequency component of at least one adjacent field and the vertical low frequency component of the current field, motion compensated by the detected motion vector. Is used as an interpolation signal, it is possible to prevent the image quality of a progressive scan-converted image from deteriorating when the motion vector is incorrect.

Further, by extracting vertical low-frequency components in the field based on the determination of the correlation between lines, it is possible to improve the moving image quality.

[0031]

Embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a progressive scan conversion circuit according to a first embodiment of the present invention. The same components as those in FIG. 8 are described with the same reference numerals.

FIG. 1 differs from FIG. 8 in that a vertical band-pass for extracting a vertical high-frequency component of at least one adjacent field, for example, the previous field, which has been motion-compensated, is provided between the motion compensation interpolation circuit 103 and the synthesis circuit 106. Filter (V-B
PF) 105, and outputs, as a scanning line interpolation signal, a signal obtained by adding the vertical high frequency component of the previous field subjected to motion compensation and the vertical low frequency component of the current field generated by the intra-field interpolation circuit 104 by the synthesis circuit 106. It is configured so that

In FIG. 1, an interlaced scanning video signal is input to an input terminal 101, and a motion vector detecting circuit 102,
It is input to the motion compensation interpolation circuit 103 and the intra-field interpolation circuit 104. The motion vector detection circuit 102 detects the motion of an image between fields as a motion vector MV.

The motion compensation interpolation circuit 103 performs motion compensation by moving the image of at least one adjacent field, for example, the image of the previous field horizontally and vertically by the motion vector MV, and performs the V-BPF 105 on the image of the motion-compensated previous field.
To supply. The pass characteristics of the V-BPF 105 at this time will be described with reference to FIG. 2. The characteristics of the V-BPF 105 are as follows. For the previous field f1, the scanning line at the center of the motion vector is YA, and the scanning line located on one line is YF. , The scanning line located one line below is defined as YG, and the V-BPF 105
Is an output signal of the V-BPF 105 = (− YF + 2YA−YG) / 4 (1). Equation (1) includes a vertical high frequency component.

Returning to FIG. 1, the output signal of the V-BPF 105 is added by the synthesizing circuit 106 according to the output signal of the intra-field interpolation circuit 104 and the motion vector MV, so that the processing range of the digital signal does not overflow or underflow. After that, the signal is input to the double speed conversion circuit 107. The input signal supplied to the input terminal 101 and the scanning line interpolation signal output from the synthesizing circuit 106 are input to the double speed conversion circuit 107, and the double speed converted sequential scanning signal is output to the output terminal 108.

FIG. 3 is a block diagram showing an example of the intra-field interpolation circuit 104. The intra-field interpolation circuit 104 shown in FIG. 3 generates an average signal of the adjacent upper and lower scanning lines YD and YE for the current field f2 (see FIG. 2).

The intra-field interpolation circuit 104 shown in FIG.
It comprises a 1H delay circuit 201, an adder 202, and a coefficient unit 203.

An input video signal from an input terminal 101 is input to a 1H delay circuit 201 and an adder 202. The scanning line located one line above is YD, and the scanning line located one line below is YE
In the adder 202, the signal of the previous line, that is, 1H
The delayed signal YD and the signal of the current line, that is, the non-delayed signal YE are added and halved by the coefficient unit 203, and the output signal of the intra-field interpolation circuit 104 becomes the output signal of the intra-field interpolation circuit 104 = ( YD + YE) / 2 (2).

Therefore, in the synthesizing circuit 106 shown in FIG. 1, as the scanning line interpolation signal, the scanning line interpolation signal = [the output signal of the V-BPF 105] + [the output signal of the intra-field interpolation circuit 104] = (-YF + 2YA-YG) / 4 + (YD + YE) / 2 (3) The signal of (3) is subjected to overflow and underflow processing and output. This signal contains a vertical high-frequency component, and the vertical resolution of a moving image is improved.

According to the present invention, in the example of the horizontal movement shown in FIG. 9, the output signal of the motion compensation interpolation circuit 103 is shown in FIG.
The video signal of the area (the block having the highest correlation in the previous field) enclosed by the broken line in FIG. Then, the vertical high frequency component is extracted by the next stage V-BPF1O5. Since there is no vertical high frequency component due to a rectangular line segment (horizontal line) as a background within the broken line in the previous field of FIG. 9A, the hatched portion in the current field of FIG. The video signal of (a) is not interpolated at all, and the image quality does not deteriorate. On the other hand, with respect to the circular portion that is the foreground, the vertical high frequency component by the circular line segment (arc line) within the broken line in FIG. 9A is motion-compensated, so that the vertical resolution can be improved.

Further, according to the present invention, FIG.
In the example of the rotation shown in (b), since the vertical high frequency component of the previous field subjected to motion compensation is added to the intra-field interpolation signal in the current field, the image obtained by sequentially performing the scan line conversion is shown in FIG. As shown in (), the horizontal edge portion becomes an image in which the scanning line is interpolated (whisker-like horizontal lines are interpolated and displayed). However, the deterioration of the image is extremely small as compared with the conventional processing.

FIG. 4 is a block diagram showing a progressive scan conversion circuit according to a second embodiment of the present invention. The same components as those in FIG. 1 will be described with the same reference numerals.

FIG. 4 differs from FIG. 1 in that the motion compensation interpolation circuit 103a performs motion compensation on the video signal of the previous field and the video signal of the next field in accordance with the detected motion vector MV, and outputs the result. Motion compensation interpolation circuit 10
A vertical band-pass filter (V-BPF) 10 for extracting a vertical high-frequency component of the video signal of the previous field, motion-compensated in 3a
5a and a vertical band-pass filter (V-BPF) 105b for extracting a vertical high-frequency component of the video signal of the field after motion compensation by the motion compensation interpolation circuit 103a. By combining the high-frequency component and the vertical low-frequency component of the current field generated by the intra-field interpolation circuit 104 according to the motion vector MV in the synthesis circuit 106a, a signal is output as a scanning line interpolation signal. is there.

In FIG. 4, an interlaced scanning video signal is input to an input terminal 101, and a motion vector detecting circuit 102,
The motion compensation interpolation circuit 103a and the intra-field interpolation circuit 104 are input. The motion vector detection circuit 102 detects the motion of an image between fields as a motion vector MV.

The motion compensation interpolation circuit 103a performs motion compensation by moving the image of at least one adjacent field, for example, the previous field and the next field horizontally and vertically by the motion vector MV, and performs the motion compensation of the previous field and the next field. Images of V-BPF 105a and V-BP
Supply to F105b. FIG. 2 shows the V-BPF 105 at this time.
a, 105b, the V-BPF1
The characteristic of the output signal of the V-BPF 105a is that the scanning line which is the center of the motion vector for the previous field f1 is YA, the scanning line located one line above is YF, and the scanning line located one line below is YG. The output signal of the V-BPF 105a = (− YF + 2YA−YG) / 8 (4).

The characteristics of the V-BPF 105b are as follows. For the subsequent field f3, the scanning line that is the center of the motion vector is YB, the scanning line located one line above is YH, and the scanning line located one line below is YI. , V-BPF105b
Is an output signal of the V-BPF 105b = (− YH + 2YB−YI) / 8 (5). Equations (4) and (5) include a vertical high frequency component.

Returning to FIG. 4, the output signals of the V-BPFs 105a and 105b are output from the combining circuit 106a to the intra-field interpolation circuit 104.
Are added according to the motion signal MV and the motion vector MV, and the digital signal is processed so that the processing range of the digital signal does not overflow or underflow.
The input signal supplied to the input terminal 101 and the scanning line interpolation signal output from the synthesis circuit 106a are input to the double speed conversion circuit 107, and the double speed converted sequential scanning signal is output to the output terminal 108.
Is output to

Here, if the circuit shown in FIG. 3 is used as the intra-field interpolation circuit 104, the synthesizing circuit 106a in FIG. 4 uses the following as the scanning line interpolation signal: scanning line interpolation signal = [output signal of V-BPF 105a] + [Output signal of V-BPF 105b] + [Output signal of inter-field interpolation circuit 104] = (-YF + 2YA-YG) / 8 + (-YH + 2YB-YI) / 8 + (YD + YE) / 2 (6) Outputs signals after overflow and underflow processing. This signal contains a vertical high-frequency component, and the vertical resolution of a moving image is improved.

FIG. 5 is a block diagram showing another example of the intra-field interpolation circuit 104. The intra-field interpolation circuit 104 shown in FIG. 5 generates a vertical low-frequency component based on the correlation between lines in the current field.

The intra-field interpolation circuit 104 shown in FIG.
1H delay circuit 301, adder 302, low-pass filters (LPF) 303 and 304, interpolation signal generation circuit 305, correlation determination circuit 306, selector 307, and high-pass filter (H
PF) 308 and an adder 309.

The input video signal from the input terminal 101 is input to a 1H delay circuit 301 and a low-pass filter (LPF) 303. The interpolation signal generation circuit 305 and the correlation determination circuit 306 include:
The output signal of the LPF 303 and the output signal of the 1H delay circuit 301 through the LPF 304 are input. The interpolation signal generation circuit 305 receives the low-frequency component of the 1H delay signal, which is the previous scanning line signal, and the low-frequency component of the non-delay signal, which is the current scanning line signal, as shown in FIG. A pair of pixel values (pixel value pair) corresponding to each other between the upper and lower lines is generated. The correlation determination circuit 306 performs a correlation determination of a pattern from the horizontal low-frequency components of the two scanning line signals, selects a pixel value pair from the interpolation signal generation circuit 305 by the selector 307 based on the pattern correlation, and determines the pair value thereof. Average and output. Then, the addition circuit 309 adds the selected output of the selector 307 to the horizontal high-frequency component of the sum signal between the upper and lower lines by the adder 302 and the high-pass filter (HPF) 308 and outputs the result.

Next, the operation of FIG. 5 will be described with reference to FIG. The correlation determination circuit 306 calculates the correlation between the pixels based on the pixel values A-3 to A + 3 of the horizontal low band component on the scanning line 1H before and the pixel values B-3 to B + 3 of the horizontal low band component on the current scanning line. Make a decision. The magnitude of the difference between the pixel value pairs (A-1, B + 1) and (A + 1, B-1) is compared, and | A-1-B + 1 |> | A + 1-B-1 | (7) | A0 -B0 |> | A + 1-B-1 | (8) If the equations (7) and (8) hold, the pixel value pairs (A + 1, B-1) and (A) +2, B-2).

| A + 1-B-1 |> | A + 2-B-2 | (9) | A + 2-B-2 |> | A + 3-B-3 | (10) If so, the pixel value pair (A + 2, B-2) is determined as the most correlated pixel value pair. If equation (8) does not hold, and | A0-B0 | ≤ | A + 1-B-1 | (11), the pixel value pair (A0, B0) is determined as the most correlated pixel value pair. Is done.

In the example of FIG. 6, the pixel value pair (A + 2, B-2) is determined as the most correlated pixel value pair, and (A + 2 + B-2) / 2 is selectively output by the selector 307. You. An adder 309 adds the horizontal high frequency component of (A0 + B0) / 2 and outputs the result.

In the output signal obtained by the intra-field interpolation circuit 104 of FIG. 5, diagonal lines are displayed more smoothly than when interpolation is performed from a simple pair of upper and lower pixels. That is, depending on the direction in which the picture changes, the aliasing component of the interlaced signal can be reduced as compared with the case where the pattern is simply generated from the upper and lower pixel values.

The vertical low-frequency component generated by the intra-field interpolation circuit 104 of FIG. 5 and the vertical high-frequency component of the previous field which has been motion-compensated by the motion vector detection described in FIG.
The double-speed conversion circuit 1
It is possible to improve the image quality of the image that is sequentially scan-converted in 07.

The vertical low-frequency component generated by the intra-field interpolation circuit 104 of FIG. 5 and the vertical high-frequency components of the previous field and the subsequent field which have been motion-compensated by the motion vector detection described in FIG. By the addition in the synthesizing circuit 106a, the image quality of the image sequentially scanned and converted by the double speed conversion circuit 107 can be improved.

Also, in the above-described embodiment, a combination of a plurality of digital circuits has been described.
Video DSP that performs programmable digital signal processing
(Digital Signal Processor) can perform the same scanning line interpolation processing.

[0059]

As described above, according to the present invention, the vertical interpolation component of the moving image is generated by adding the vertical high frequency component of the previous field and the vertical low frequency component of the current field which have been motion compensated. It is possible to improve the high frequency component and reduce the aliasing component due to the interlace signal. Also,
Even when the magnitude and direction of the motion are different between the background and the foreground, or when the motion of the object is rotating, the deterioration of the image quality can be prevented by limiting the component by the motion compensation to the vertical high frequency component. In other words, it is possible to increase the frequency of motion compensation in the vertical high frequency region.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a progressive scan conversion circuit according to an embodiment of the present invention.

FIG. 2 is a view for explaining the operation of the embodiment in FIG. 1;

FIG. 3 is a block diagram showing an example of an intra-field interpolation circuit.

FIG. 4 is a block diagram showing a progressive scan conversion circuit according to another embodiment of the present invention.

FIG. 5 is a block diagram showing another example of the intra-field interpolation circuit.

FIG. 6 is a view for explaining the operation of FIG. 5;

FIG. 7 is a diagram illustrating a method for detecting a motion vector.

FIG. 8 is a block diagram showing a conventional progressive scan conversion circuit.

FIG. 9 is a view for explaining the drawbacks of the conventional example of FIG. 8;

FIG. 10 is a diagram illustrating a conventional technique.

FIG. 11 is a view for explaining the drawbacks of the conventional example of FIG. 10;

FIG. 12 is a view for explaining a drawback of the conventional sequential conversion.

[Explanation of symbols]

 102: motion vector detection circuit 103, 103a: motion compensation interpolation circuit 104: intra-field interpolation circuit 105, 105a, 105b ... vertical band-pass filter 106, 106a ... synthesis circuit

Claims (5)

[Claims] 1. A progressive scan conversion circuit for converting an interlaced video signal into a progressive scan video signal, comprising: a motion vector detection circuit for detecting a motion of an image between frames or fields of a video signal as a motion vector; A motion compensation interpolator for motion-compensating and outputting a video signal of at least one adjacent field in accordance with the obtained motion vector; a circuit for extracting a vertical high-frequency component of the motion-compensated field; A progressive scan conversion circuit, comprising: a circuit for generating a component; and a synthesis circuit for adding the previously allowed vertical high frequency component and the vertical low frequency component and outputting the result as an interpolation signal. 2. The motion compensation interpolation circuit according to claim 1, wherein the motion compensation interpolation circuit performs motion compensation on a video signal of a previous field and outputs the video signal. The bandpass characteristic of (−YF + 2YA−YG) / 4 is defined assuming that the scanning line that is the center of the motion vector for the previous field is YA, the scanning line located one line above is YF, and the scanning line located one line below is YG. The circuit for generating the vertical low-frequency component of the current field generates an average signal of adjacent upper and lower scanning lines with respect to the current field. 2. The progressive scan conversion according to claim 1, wherein the scanning line located one line below has a low-pass characteristic of (YD + YE) / 2 as YE. Road. 3. The motion compensation interpolation circuit according to claim 1, wherein:
The video signal of the next field is motion-compensated and output, and the circuit for extracting the vertical high-frequency component of at least one adjacent field which has been motion-compensated includes a scanning line serving as a center of a motion vector for the previous field. Is YA, the scanning line located one line above is YF, and the scanning line located one line below is YG, the bandpass characteristic is (−YF + 2YA−YG) / 8, and the center of the motion vector is The scanning line having a band-pass characteristic of (−YH + 2YB−YI) / 8 is defined as YB, a scanning line located one line above the scanning line is YH, and a scanning line located one line below the scanning line is YI. The circuit for generating the vertical low-frequency component of the field generates an average signal of adjacent upper and lower scanning lines for the current field. Progressive scanning conversion circuit according to claim 1, wherein the scan lines of the scan lines located under the YD, 1 line and has a (YD + YE) / 2 of the low-pass characteristic as YE located. 4. A progressive scan conversion circuit for converting an interlaced video signal into a progressive scan video signal, comprising: a motion vector detection circuit for detecting a motion of an image between frames or fields of a video signal as a motion vector; A motion compensation interpolator for motion-compensating and outputting a video signal of at least one adjacent field in accordance with the obtained motion vector; a circuit for extracting a vertical high-frequency component of the motion-compensated field; A progressive scan conversion circuit comprising: a circuit that generates a vertical low frequency component based on the correlation determination of: and a synthesis circuit that adds the vertical high frequency component and the vertical low frequency component and outputs the result as an interpolation signal. . 5. The motion-compensated at least one adjacent one
The circuit for extracting the vertical high-frequency component of one field includes a scanning line that is the center of the motion vector for the previous field as YA, a scanning line located one line above as YF, and a scanning line located below one line as YG. (−YF + 2YA−YG) / 4, and the circuit that generates the vertical low-frequency component based on the correlation determination between the lines in the current field includes: Judgment of correlation between pixels from the pixel values A-3 to A + 3 of the horizontal low band component on the scanning line on one line and the pixel values B-3 to B + 3 of the horizontal low band component on the scanning line one line below And pixel value pairs (A-1, B + 1), (A + 1, B
A-1−B + 1 |> | A + 1−B−1 | (a) | A0−B0 |> | A + 1−B-1 | (b) If the expressions (a) and (b) hold, the pixel value pair (A + 1,
B-1) and (A + 2, B-2) are compared, and | A + 1-B-1 |> | A + 2-B-2 | | A + 2-B-2 | > | A + 3−B-3 |, the pixel value pair (A + 2, B-2) is determined as the most correlated pixel value pair, and equation (b) does not hold, and | A0 -B0 | ≤ | A + 1-B-1 |, the pixel value pair (A0, B0) is determined as the most correlated pixel value pair, and the pixel value determined as the most correlated pixel value pair Pair (A + 2, B-2) or (A0, B
0), (A + 2 + B-2) / 2 or (A0 + B0)
) / 2 is generated, and the generated output and (A0 + B0)
5. The progressive scan conversion circuit according to claim 4, wherein the output is obtained by adding a horizontal high-frequency component of / 2.